FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2000153091> ?p ?o ?g. }

• W2000153091 endingPage "291" @default.
• W2000153091 startingPage "282" @default.
• W2000153091 abstract "•Whole-genome sequencing is used for forensic epidemiology.•Big data can transform forensic epidemiology.•Clustering, biases, wildlife reservoirs, and emerging infections can all be addressed.•Phylodynamics approaches to integrate epidemiological and evolutionary data have been highly successful but still face scientific challenges. In epidemiology, the identification of ‘who infected whom’ allows us to quantify key characteristics such as incubation periods, heterogeneity in transmission rates, duration of infectiousness, and the existence of high-risk groups. Although invaluable, the existence of many plausible infection pathways makes this difficult, and epidemiological contact tracing either uncertain, logistically prohibitive, or both. The recent advent of next-generation sequencing technology allows the identification of traceable differences in the pathogen genome that are transforming our ability to understand high-resolution disease transmission, sometimes even down to the host-to-host scale. We review recent examples of the use of pathogen whole-genome sequencing for the purpose of forensic tracing of transmission pathways, focusing on the particular problems where evolutionary dynamics must be supplemented by epidemiological information on the most likely timing of events as well as possible transmission pathways. We also discuss potential pitfalls in the over-interpretation of these data, and highlight the manner in which a confluence of this technology with sophisticated mathematical and statistical approaches has the potential to produce a paradigm shift in our understanding of infectious disease transmission and control. In epidemiology, the identification of ‘who infected whom’ allows us to quantify key characteristics such as incubation periods, heterogeneity in transmission rates, duration of infectiousness, and the existence of high-risk groups. Although invaluable, the existence of many plausible infection pathways makes this difficult, and epidemiological contact tracing either uncertain, logistically prohibitive, or both. The recent advent of next-generation sequencing technology allows the identification of traceable differences in the pathogen genome that are transforming our ability to understand high-resolution disease transmission, sometimes even down to the host-to-host scale. We review recent examples of the use of pathogen whole-genome sequencing for the purpose of forensic tracing of transmission pathways, focusing on the particular problems where evolutionary dynamics must be supplemented by epidemiological information on the most likely timing of events as well as possible transmission pathways. We also discuss potential pitfalls in the over-interpretation of these data, and highlight the manner in which a confluence of this technology with sophisticated mathematical and statistical approaches has the potential to produce a paradigm shift in our understanding of infectious disease transmission and control. Identifying pathways of infectious disease transmission can reveal likely points of control and predict future directions of spread. In combination with mathematical models (see Glossary) they can be used to predict the outcomes of alternative control methods. Central to this is epidemiological tracing to identify ‘who infected whom’, a crucial component of what is known as forensic epidemiology. Unfortunately, tracing is often made difficult by the effort required and the considerable uncertainties in the possible sources of infection and timings of events. Contact patterns can sometimes be inferred from spatiotemporal proximity, particularly where the host populations are sessile and with short-range contacts {e.g., foot-and-mouth disease (FMD) on farms [1Ferguson N.M. et al.The foot-and-mouth epidemic in Great Britain: pattern of spread and impact of interventions.Science. 2001; 292: 1155-1160Crossref PubMed Scopus (539) Google Scholar], citrus canker in fruit trees [2Parnell S. et al.Optimal strategies for the eradication of asiatic citrus canker in heterogeneous host landscapes.Phytopathology. 2009; 99: 1370-1376Crossref PubMed Scopus (38) Google Scholar], rabies in domestic dogs [3Hampson K. et al.Transmission dynamics and prospects for the elimination of canine rabies.PLoS Biol. 2009; 7: 462-471Crossref Scopus (326) Google Scholar], and hospital infections [4Harris S.R. et al.Evolution of MRSA during hospital transmission and intercontinental spread.Science. 2010; 327: 469-474Crossref PubMed Scopus (879) Google Scholar]} or through the identification of relevant risk factors (e.g., needle-sharing or sexual contact for HIV transmission). However, even in these cases the difficulty of identifying the most relevant routes and means of contact limits our ability to characterize the underlying transmission processes. Antigenic or genetic characterization [e.g., serotyping or multi-locus sequence typing (MLST)] of pathogens is an alternative approach to identifying groups of individuals with closely related infections [5Maiden M.C. et al.Multilocus sequence typing: a portable approach to the identification of clones within populations of pathogenic microorganisms.Proc. Natl. Acad. Sci. U.S.A. 1998; 95: 3140-3145Crossref PubMed Scopus (2919) Google Scholar]. Until recently these approaches lacked the resolution for characterizing direct contact. However, high-throughput sequencing (HTS) technology, together with improved ability to extract genetic material more cheaply and from smaller pathogen samples [6Koser C.U. et al.Routine use of microbial whole genome sequencing in diagnostic and public health microbiology.PLoS Pathog. 2012; 8: e1002824Crossref PubMed Scopus (384) Google Scholar, 7Dumitrescu O. et al.Present and future automation in bacteriology.Clin. Microbiol. Infect. 2011; 17: 649-650Crossref PubMed Scopus (14) Google Scholar], now allow mass-scale characterization of virtually entire genomes of whole populations of pathogens (generally referred to as whole-genome sequencing or WGS). This technology typically offers orders of magnitude better resolution compared to earlier typing methods [8Roetzer A. et al.Whole genome sequencing versus traditional genotyping for investigation of a Mycobacterium tuberculosis outbreak: a longitudinal molecular epidemiological study.PLoS Med. 2013; 10: e1001387Crossref PubMed Scopus (366) Google Scholar]. In addition, the increased availability of dense data characterizing the substrate population (e.g., identification of individuals, social groupings, contacts between groups, spatial organization, species compositions etc., and referred to here as denominator data) [9Kao R.R. et al.Disease dynamics over very different time-scales: foot-and-mouth disease and scrapie on the network of livestock movements in the UK.J. R. Soc. Interface. 2007; 4: 907-916Crossref PubMed Scopus (132) Google Scholar, 10Bajardi P. et al.Optimizing surveillance for livestock disease spreading through animal movements.J. R. Soc. Interface. 2012; 9: 2814-2825Crossref PubMed Scopus (103) Google Scholar], and the development of powerful computational and analytical tools to organize and interpret large datasets, broadens the potential for application of such data to high-resolution epidemiological problems. Although their usage on a large scale is in its infancy, they share many properties with ‘big data’ problems in other systems: (i) although highly variable in size, big datasets are typically an order of magnitude or greater larger than what had previously been available, (ii) the proportion and coverage of data on the susceptible population of interest that is captured in the datasets are high, and (iii) the variety of data being captured is extensive. The opportunities presented by big data based on WGS are potentially paradigm-shifting, with existing smaller-scale studies [11Morelli M.J. et al.A Bayesian inference framework to reconstruct transmission trees using epidemiological and genetic data.PLoS Comput. Biol. 2012; 8: e1002768Crossref PubMed Scopus (87) Google Scholar, 12Ypma R.J. et al.Unravelling transmission trees of infectious diseases by combining genetic and epidemiological data.Proc. Biol. Sci. 2012; 279: 444-450Crossref PubMed Scopus (107) Google Scholar] hinting at what might be possible with very large datasets. Crucial to this is the integration of non-WGS data into analyses identifying epidemiological pathways because this can lead to a considerable refinement of our understanding of transmission. Although this is often conducted descriptively, ‘epidemiological’ frameworks are being developed that naturally incorporate genetic data with both denominator data and additional information on the transmission of the pathogen across the affected population. In the remainder of this review we shall consider the role that WGS can play in enhancing our understanding of fine-scale epidemiological contact. We shall highlight the pitfalls that arise if there are multiple likely transmission routes for every true transmission route, and where there are differences between observed phylogenies and transmission networks, including the difficulties of inferring the epidemiological dynamics of multi-host pathogens and emerging infections. The majority of mutations for any pathogen will be subject to strong purifying selection, with a small minority being subject to positive selection (and potentially a problematic source of homoplasy). This still leaves substantial numbers of neutral or ‘nearly neutral’ mutations (i.e., sites subject to only weak selection) [13Bhatt S. et al.The genomic rate of molecular adaptation of the human influenza A virus.Mol. Biol. Evol. 2011; 28: 2443-2451Crossref PubMed Scopus (110) Google Scholar]. Although such nearly neutral variation may be selected out over longer time scales [14Morelli M.J. et al.Evolution of foot-and-mouth disease virus intra-sample sequence diversity during serial transmission in bovine hosts.Vet. Res. 2013; 44: 12Crossref PubMed Scopus (51) Google Scholar, 15Nelson M.I. Holmes E.C. The evolution of epidemic influenza.Nat. Rev. Genet. 2007; 8: 196-205Crossref PubMed Scopus (404) Google Scholar], over shorter time scales such as a single epidemic they can be useful markers of pathogen genealogy, provided that phenotypic effects [16Lee R.T. et al.All that glitters is not gold – founder effects complicate associations of flu mutations to disease severity.Virol. J. 2010; 7: 297Crossref PubMed Scopus (21) Google Scholar] are minimal. These mutations will not necessarily be synonymous because there may be constraints imposed by genetic structure (e.g., RNA secondary structure) and overlapping reading frames (i.e., a synonymous mutation on one frame can be nonsynonymous and selected against in the other) [17Holmes E.C. The evolutionary genetics of emerging viruses.Annu. Rev. Ecol. Evol. Syst. 2009; 40: 353-372Crossref Scopus (137) Google Scholar]. Polymorphisms in sets of sequences can be compromised by technical issues, including errors in sequencing and bioinformatics, resulting in missed or artefactually added mutations), by reassortment in segmented genomes such as in influenza viruses, and by recombination in non-segmented genomes such as those of retroviruses or bacteria [18Croucher N.J. et al.Bacterial genomes in epidemiology – present and future.Philos. Trans. R. Soc. Lond. B: Biol. Sci. 2013; 368: 20120202Crossref PubMed Scopus (41) Google Scholar, 19Marttinen P. et al.Detection of recombination events in bacterial genomes from large population samples.Nucleic Acids Res. 2012; 40: e6Crossref PubMed Scopus (151) Google Scholar]. All amplification steps can introduce errors, and the more amplification that is required the more likely that errors will be introduced. The number and nature of the artefacts introduced will therefore depend on the size of the original genetic sample, the laboratory protocols used (including the reagents used to process a given sample), the sequencing technology, and also the analytical tools used, with a lack of agreed quality-control protocols providing an additional layer of uncertainty. The nature of the pathogen itself is also important, with RNA viruses requiring error-prone reverse transcription [20Sanjuan R. et al.Viral mutation rates.J. Virol. 2010; 84: 9733-9748Crossref PubMed Scopus (827) Google Scholar]. Such errors carry identifiable signatures; for example artefacts are more likely to be random and appear at low frequency across replicates, unlike the ‘true’ mutations because these should almost always appear. Methods to identify and minimize these errors are being identified [21Lou D.I. et al.High-throughput DNA sequencing errors are reduced by orders of magnitude using circle sequencing.Proc. Natl. Acad. Sci. U.S.A. 2013; 110: 19872-19877Crossref PubMed Scopus (198) Google Scholar, 22Beerenwinkel N. et al.Challenges and opportunities in estimating viral genetic diversity from next-generation sequencing data.Front. Microbiol. 2012; 3: 329Crossref PubMed Scopus (184) Google Scholar]. In the absence of horizontal genetic transfer the genetic distance between sequenced pathogens is usually positively correlated with the number of transmission links between individuals. For tracing contact there would ideally be a unique sequence that is shared by the entire within-host population but, immediately upon transmission, would acquire at least one distinguishing mutation. Unfortunately, such a pathogen does not exist, resulting in multiple complications (Box 1). Mutation rates will vary, sometimes considerably [23Bryant J.M. et al.Inferring patient to patient transmission of Mycobacterium tuberculosis from whole genome sequencing data.BMC Infect. Dis. 2013; 13: 110Crossref PubMed Scopus (149) Google Scholar], as will the times between consecutive transmission events (referred to here as generation times). Further complications arise should epidemiological processes influence evolutionary rates. Examples include the potential role of duration of latency in the mutation rates of Mycobacterium tuberculosis in humans [24Walker T.M. et al.Whole-genome sequencing to delineate Mycobacterium tuberculosis outbreaks: a retrospective observational study.Lancet Infect. Dis. 2013; 13: 137-146Abstract Full Text Full Text PDF PubMed Scopus (640) Google Scholar] and evidence that tropical and subtropical climates accelerate the evolution of bat rabies virus [25Streicker D.G. et al.Rates of viral evolution are linked to host geography in bat rabies.PLoS Pathog. 2012; 8: e1002720Crossref PubMed Scopus (72) Google Scholar]. When mutation rates compared to generation times are low (which we shall call the relative mutation rate), WGS data can provide insufficient genetic signal. For instance, during explosive outbreaks of acute or hyperacute viral infections (i.e., influenza and norovirus), the relative mutation rate might be very low, and consensus sequences among samples from different but closely-related infected individuals may be identical. Such situations may be resolved by using within-host genetic variation to infer properties of between-host transmission [26Stack J.C. et al.Inferring the inter-host transmission of influenza A virus using patterns of intra-host genetic variation.Proc. Biol. Sci. 2013; 280: 20122173Crossref PubMed Scopus (36) Google Scholar], as could be achieved by examining the presence of minor allele variants shared by different hosts.Box 1Phylogenies and transmission treesAt coarse spatial and temporal resolutions the evolutionary relationships between pathogen genes will reflect their epidemiological relationships. Pathogens that are closely related epidemiologically will also be those most closely related to each other genetically. However, it is well known [71Pybus O.G. Rambaut A. Evolutionary analysis of the dynamics of viral infectious disease.Nat. Rev. Genet. 2009; 10: 540-550Crossref PubMed Scopus (418) Google Scholar] that at finer space–time resolution the particular details of the epidemiological process can begin to decouple the transmission tree from the genealogy. In Figure I, filled circles represent genomes sampled from four particular host individuals (hosts labeled A–D and colored red, blue, black, and green respectively). Unfilled circles represent genomes that were not sampled. Circles immediately adjacent to one another are one mutation different to each other. The color of the line indicates which host the different genomes were in. In (i), the pathogen genome sampled is the genome that was transmitted. We can therefore deduce that the most likely transmission scenario is that A infected B, B infected C, and C infected D. However, if transmission occurred sometime before the time of sampling, such that additional mutations were subsequently incurred, either because the infectious period is long, or the mutation rate very high, then we inevitably become less certain which genome was in which host, and consequently several different transmission trees become consistent with the genetic data. It may be that A infected B, B infected C, and C infected D (ii), but the genetic data are equally consistent with a scenario in which D infected A, B, and C (iii), or indeed several alternative explanations (not shown). It is in these situations that the integration of additional data on the timing and the contact process becomes important for inferring the most likely tree. There are other reasons why the transmission tree and phylogeny may be different – for example there may be insufficient genetic information to distinguish between pathogen from different hosts, or recombination or homoplasy may complicate the relationship between the two. At coarse spatial and temporal resolutions the evolutionary relationships between pathogen genes will reflect their epidemiological relationships. Pathogens that are closely related epidemiologically will also be those most closely related to each other genetically. However, it is well known [71Pybus O.G. Rambaut A. Evolutionary analysis of the dynamics of viral infectious disease.Nat. Rev. Genet. 2009; 10: 540-550Crossref PubMed Scopus (418) Google Scholar] that at finer space–time resolution the particular details of the epidemiological process can begin to decouple the transmission tree from the genealogy. In Figure I, filled circles represent genomes sampled from four particular host individuals (hosts labeled A–D and colored red, blue, black, and green respectively). Unfilled circles represent genomes that were not sampled. Circles immediately adjacent to one another are one mutation different to each other. The color of the line indicates which host the different genomes were in. In (i), the pathogen genome sampled is the genome that was transmitted. We can therefore deduce that the most likely transmission scenario is that A infected B, B infected C, and C infected D. However, if transmission occurred sometime before the time of sampling, such that additional mutations were subsequently incurred, either because the infectious period is long, or the mutation rate very high, then we inevitably become less certain which genome was in which host, and consequently several different transmission trees become consistent with the genetic data. It may be that A infected B, B infected C, and C infected D (ii), but the genetic data are equally consistent with a scenario in which D infected A, B, and C (iii), or indeed several alternative explanations (not shown). It is in these situations that the integration of additional data on the timing and the contact process becomes important for inferring the most likely tree. There are other reasons why the transmission tree and phylogeny may be different – for example there may be insufficient genetic information to distinguish between pathogen from different hosts, or recombination or homoplasy may complicate the relationship between the two. The consideration of within-host dynamics has been shown to reconcile the differences between the sampled phylogeny and transmission dynamics, improving the inferred transmission tree [27Ypma R.J. et al.Relating phylogenetic trees to transmission trees of infectious disease outbreaks.Genetics. 2013; 195: 1055-1062Crossref PubMed Scopus (109) Google Scholar, 28Onnela J.P. et al.Geographic constraints on social network groups.PLoS ONE. 2011; 6: e16939Crossref PubMed Scopus (247) Google Scholar]. When relative mutation rates are high the differences between time of transmission and sample time, and within-host location of sampling compared to location at which transmission occurs, can increase the observed genetic distances between mother–daughter pairs and introduce ambiguities and biases in the available data (Box 1). For chronic infections such as HIV, hepatitis C, and tuberculosis, an individual could be diagnosed months or even years after transmission and thus the first onward transmission event may pre-date the consensus sequence [29Volz E.M. et al.Simple epidemiological dynamics explain phylogenetic clustering of HIV from patients with recent infection.PLoS Comput. Biol. 2012; 8: e1002552Crossref PubMed Scopus (78) Google Scholar]; this can also be a problem for acute-acting viruses such as FMD virus (FMDV) [30Cottam E.M. et al.Integrating genetic and epidemiological data to determine transmission pathways of foot-and-mouth disease virus.Proc. Biol. Sci. 2008; 275: 887-895Crossref PubMed Scopus (131) Google Scholar]. The problems are exacerbated where many intermediate cases may be absent from the data, although these may be alleviated by the development of temporal markers of infection or reliable indicators of change in the microbial community [31Zaas A.K. et al.A host-based RT-PCR gene expression signature to identify acute respiratory viral infection.Sci. Transl. Med. 2013; 5: 203ra126Crossref PubMed Scopus (102) Google Scholar]. Further difficulties will be incurred if we are interested in transmission processes at multiple scales because the ‘ideal’ rate would be different at each scale. Mechanistic models of infectious disease transmission (often called mathematical models) can be used to generate simulated transmission trees based on our understanding of the underlying mechanisms that drive the transmission process (Box 2). Because they are mechanistic, by altering the mechanisms in the model they can be used to predict future outcomes of ongoing epidemics ‘how big, and how long?’. It can also be used to predict the outcome of interventions, such as ‘will mass vaccination be effective, and what is the required coverage?’, ‘how many anti-viral drugs will be needed during an influenza pandemic?’, or ‘is culling or vaccination a better policy to control FMD?’. Such models are designed to provide population-level insights and are typically poor at predicting actual events at small scales because the number of possible transmission trees that result in a given observed epidemic can be very large. When there are even low levels of error and uncertainty in the available data, the accuracy of parameter estimates can be severely degraded [32Savill N.J. et al.Effect of data quality on estimates of farm infectiousness trends in the UK 2001 foot-and-mouth disease epidemic.J. R. Soc. Interface. 2007; 4: 235-241Crossref PubMed Scopus (14) Google Scholar]. However, good denominator data can substantially reduce the range of possible contacts that result in observed patterns of transmission. Big denominator data are becoming more common; densely sampled associations being recorded include daily, individual movement records for livestock [9Kao R.R. et al.Disease dynamics over very different time-scales: foot-and-mouth disease and scrapie on the network of livestock movements in the UK.J. R. Soc. Interface. 2007; 4: 907-916Crossref PubMed Scopus (132) Google Scholar, 10Bajardi P. et al.Optimizing surveillance for livestock disease spreading through animal movements.J. R. Soc. Interface. 2012; 9: 2814-2825Crossref PubMed Scopus (103) Google Scholar], and mobile phone network [33Onnela J.P. et al.Structure and tie strengths in mobile communication networks.Proc. Natl. Acad. Sci. U.S.A. 2007; 104: 7332-7336Crossref PubMed Scopus (1329) Google Scholar] and airline traffic data [34Hufnagel L. et al.Forecast and control of epidemics in a globalized world.Proc. Natl. Acad. Sci. U.S.A. 2004; 101: 15124-15129Crossref PubMed Scopus (782) Google Scholar, 35Bajardi P. et al.Human mobility networks, travel restrictions, and the global spread of 2009 H1N1 pandemic.PLoS ONE. 2011; 6: e16591Crossref PubMed Scopus (318) Google Scholar] for humans. The incorporation of big denominator data into epidemiological models is greatly aided by Bayesian statistical inference frameworks that formalize the relationships between prior knowledge and model-derived likelihood functions. The technical challenges of accomplishing this are not to be underestimated [36Chis Ster I. et al.Within-farm transmission dynamics of foot and mouth disease as revealed by the 2001 epidemic in Great Britain.Epidemics. 2012; 4: 158-169Abstract Full Text Full Text PDF PubMed Scopus (32) Google Scholar], and identifying whether the best-fit model is a good model, particularly where approximate methods have been used, is challenging [37Templeton A.R. Statistical hypothesis testing in intraspecific phylogeography: nested clade phylogeographical analysis vs. approximate Bayesian computation.Mol. Ecol. 2009; 18: 319-331Crossref PubMed Scopus (87) Google Scholar]. These problems are further complicated by the often uncertain relationship between what are often multiple putative routes of transmission and their relative importance to the transmission tree. Although many of the methods used to analyze WGS data are extensions of previous approaches, in one way, WGS provides a unique insight; its unusually dense information can change our understanding of the transmission process at the individual transmission event scale. Also unlike other sources of data, the genealogical relationships are fundamental to the transmission tree, even if the relationship between the transmission trees and the observed genealogies is imperfect (Box 1 and Figure 1). However, ambiguities and errors can be substantially reduced by combining WGS with population-level inference, thereby joining the epidemiological (individual-to-individual) and ecological (demographic, population-level) perspectives [11Morelli M.J. et al.A Bayesian inference framework to reconstruct transmission trees using epidemiological and genetic data.PLoS Comput. Biol. 2012; 8: e1002768Crossref PubMed Scopus (87) Google Scholar]. Particular problems are considered below.Box 2Bayesian model-based inferenceSpatial, temporal, and pathogen genetic information have been used in two broadly different ways to reconstruct the dynamics of epidemics. In the first, coalescent models that assume a particular population dynamic model are used to link the demography of the pathogen to its evolution; in this approach a flexible diffusion-like process can be used to estimate the rate of spatial spread of the pathogen [72Lemey P. et al.Phylogeography takes a relaxed random walk in continuous space and time.Mol. Biol. Evol. 2010; 27: 1877-1885Crossref PubMed Scopus (465) Google Scholar]. This enables estimation of several useful parameters, including those describing the pathogen demography [73Drummond A.J. et al.Estimating mutation parameters, population history and genealogy simultaneously from temporally spaced sequence data.Genetics. 2002; 161: 1307-1320PubMed Google Scholar], the diffusivity of the pathogen [74Pybus O.G. et al.Unifying the spatial epidemiology and molecular evolution of emerging epidemics.Proc. Natl. Acad. Sci. U.S.A. 2012; 109: 15066-15071Crossref PubMed Scopus (179) Google Scholar], and the molecular clock [73Drummond A.J. et al.Estimating mutation parameters, population history and genealogy simultaneously from temporally spaced sequence data.Genetics. 2002; 161: 1307-1320PubMed Google Scholar]. The method is robust to the density of sampling but, because such models are underpinned by fundamentally ecological (or demographic) processes, the estimated parameters do not have straightforward epidemiological interpretations and inferences about high-resolution epidemiological processes are not easily made [75Dearlove B. Wilson D.J. Coalescent inference for infectious disease: meta-analysis of hepatitis C. Philosophical transactions of the Royal Society of London. Series B.Biol. Sci. 2013; 368: 20120314Crossref Scopus (36) Google Scholar, 76Volz E.M. Frost S.D. Inferring the source of transmission with phylogenetic data.PLoS Comput. Biol. 2013; 9: e1003397Crossref PubMed Scopus (52) Google Scholar, 77Stadler T. et al.Birth–death skyline plot reveals temporal changes of epidemic spread in HIV and hepatitis C virus (HCV).Proc. Natl. Acad. Sci. U.S.A. 2013; 110: 228-233Crossref PubMed Scopus (305) Google Scholar]. Indeed, the more highly temporally resolved the data become the more important it is that epidemiological processes are given explicit representation if transmission is to be represented accurately, and the greater the shortcomings that a fundamentally ecological approach has, as opposed to an epidemiological one. Coalescent models can be modified to include an explicit epidemiological focus [41Volz E.M. et al.Phylodynamics of infectious disease epidemics.Genetics. 2009; 183: 1421-1430Crossref PubMed Scopus (154) Google Scholar], but they do not as yet account for the high levels of clustering that characterizes spatial spread, for example.The second approach combines explicit models of transmission with simple models of genetic drift to reconstruct transmission trees reflecting ‘who infected whom’. This approach recognizes the host population structure and the epidemiological processes that govern the interaction of host and pathogen. An epidemiologic" @default.
• W2000153091 created "2016-06-24" @default.
• W2000153091 creator A5004828941 @default.
• W2000153091 creator A5014427910 @default.
• W2000153091 creator A5059869899 @default.
• W2000153091 creator A5086188845 @default.
• W2000153091 date "2014-05-01" @default.
• W2000153091 modified "2023-10-10" @default.
• W2000153091 title "Supersize me: how whole-genome sequencing and big data are transforming epidemiology" @default.
• W2000153091 cites W1544540079 @default.
• W2000153091 cites W1749928497 @default.
• W2000153091 cites W1915699870 @default.
• W2000153091 cites W1970771372 @default.
• W2000153091 cites W1971421925 @default.
• W2000153091 cites W1980030331 @default.
• W2000153091 cites W1980099211 @default.
• W2000153091 cites W1981581556 @default.
• W2000153091 cites W1981720935 @default.
• W2000153091 cites W1981770389 @default.
• W2000153091 cites W1992616459 @default.
• W2000153091 cites W1993210385 @default.
• W2000153091 cites W1994552520 @default.
• W2000153091 cites W1998386958 @default.
• W2000153091 cites W2000859682 @default.
• W2000153091 cites W2010580288 @default.
• W2000153091 cites W2010726107 @default.
• W2000153091 cites W2016936944 @default.
• W2000153091 cites W2018640297 @default.
• W2000153091 cites W2023791005 @default.
• W2000153091 cites W2029758788 @default.
• W2000153091 cites W2030540141 @default.
• W2000153091 cites W2035738809 @default.
• W2000153091 cites W2041753380 @default.
• W2000153091 cites W2046702811 @default.
• W2000153091 cites W2054034186 @default.
• W2000153091 cites W2057434754 @default.
• W2000153091 cites W2059041686 @default.
• W2000153091 cites W2070521125 @default.
• W2000153091 cites W2074329989 @default.
• W2000153091 cites W2074607858 @default.
• W2000153091 cites W2080618984 @default.
• W2000153091 cites W2080723129 @default.
• W2000153091 cites W2088267428 @default.
• W2000153091 cites W2091264144 @default.
• W2000153091 cites W2092772380 @default.
• W2000153091 cites W2096091969 @default.
• W2000153091 cites W2098062910 @default.
• W2000153091 cites W2100346081 @default.
• W2000153091 cites W2100776472 @default.
• W2000153091 cites W2101091363 @default.
• W2000153091 cites W2105226322 @default.
• W2000153091 cites W2105245948 @default.
• W2000153091 cites W2106476480 @default.
• W2000153091 cites W2108857546 @default.
• W2000153091 cites W2110994363 @default.
• W2000153091 cites W2112515389 @default.
• W2000153091 cites W2114766653 @default.
• W2000153091 cites W2116204383 @default.
• W2000153091 cites W2116870694 @default.
• W2000153091 cites W2121009566 @default.
• W2000153091 cites W2122467784 @default.
• W2000153091 cites W2122919191 @default.
• W2000153091 cites W2127435093 @default.
• W2000153091 cites W2131071075 @default.
• W2000153091 cites W2131158924 @default.
• W2000153091 cites W2131871678 @default.
• W2000153091 cites W2131994108 @default.
• W2000153091 cites W2133488722 @default.
• W2000153091 cites W2133979592 @default.
• W2000153091 cites W2137553994 @default.
• W2000153091 cites W2139166200 @default.
• W2000153091 cites W2140885215 @default.
• W2000153091 cites W2141113219 @default.
• W2000153091 cites W2142503351 @default.
• W2000153091 cites W2145310559 @default.
• W2000153091 cites W2145605402 @default.
• W2000153091 cites W2146434926 @default.
• W2000153091 cites W2149753281 @default.
• W2000153091 cites W2151265018 @default.
• W2000153091 cites W2160253520 @default.
• W2000153091 cites W2166236011 @default.
• W2000153091 cites W2166867592 @default.
• W2000153091 cites W2168358002 @default.
• W2000153091 cites W2170191171 @default.
• W2000153091 cites W2319051260 @default.
• W2000153091 doi "https://doi.org/10.1016/j.tim.2014.02.011" @default.
• W2000153091 hasPubMedCentralId "https://www.ncbi.nlm.nih.gov/pmc/articles/7125769" @default.
• W2000153091 hasPubMedId "https://pubmed.ncbi.nlm.nih.gov/24661923" @default.
• W2000153091 hasPublicationYear "2014" @default.
• W2000153091 type Work @default.
• W2000153091 sameAs 2000153091 @default.
• W2000153091 citedByCount "114" @default.
• W2000153091 countsByYear W20001530912014 @default.
• W2000153091 countsByYear W20001530912015 @default.
• W2000153091 countsByYear W20001530912016 @default.
• W2000153091 countsByYear W20001530912017 @default.
• W2000153091 countsByYear W20001530912018 @default.
• W2000153091 countsByYear W20001530912019 @default.

• next